Sgarbossa EB, Pinski SL, Barbagelata A, et al. Electrocardiographic diagnosis of evolving acute myocardial infarction in the presence of left bundle-branch block. N Engl J Med. 1996 Feb 22;334(8):481-7.
What we already know about the topic: When a patient presents with acute myocardial infarction (AMI), time is myocardium – that is, the patient should receive coronary reperfusion therapy (catheterization or thrombolysis) as soon as possible. The presence of a left bundle branch block (LBBB) on ECG makes the ECG more difficult to analyze for signs of ischemia.
Why this study is important: This study attempts to identify ECG characteristics consistent with AMI in patients with acute chest pain whose ECGs also include a LBBB. Diagnosing AMI faster and more accurately in these patients gives them a better chance for a good outcome.
Brief overview of the study: First, the authors used a derivation sample to determine criteria for the diagnosis of AMI in the presence of LBBB. Their study group consisted of the North American patients in the GUSTO-1 trial who had both enzyme-confirmed AMI and LBBB on ECG. The control group consisted of randomly selected patients with stable coronary artery disease and LBBB from the Duke Data-bank for Cardiovascular Disease. Each ECG was interpreted for signs of myocardial injury by one of four cardiologists, who were blinded to the patients’ diagnoses. Signs of myocardial injury were then arranged in different combinations to maximize specificity for diagnosis of AMI. The most specific set of criteria were: 1) ST segment elevation > 1 mm and concordant with QRS complex (5 points), 2) ST-segment depression > 1 mm in lead V1, V2, or V3 (3 points), or 3) ST-segment elevation > 5 mm and discordant with QRS complex (2 points). A score of > 3 points was 90% specific for diagnosis of AMI. The criteria were then tested in a validation sample, which consisted of 22 patients with enzyme-proven AMI and 23 patients with unstable angina. The index score correctly classified 67% of patients in the validation study (84% had been correctly classified in the derivation sample).
Limitations: The study did not attempt to distinguish between new and old LBBBs. The authors postulated at the time that in clinical practice, an old ECG would rarely be rapidly accessible anyway. With the near-complete use of electronic medical record systems, this is no longer the case. The ECG readers in this study were all cardiologists, not emergency physicians. Finally, choosing 5 mm for the necessary amount of discordant elevation is somewhat arbitrary. The modified Sgarbossa criteria (Smith et al, 2012) are unweighted, and attempt to improve diagnostic utility by replacing criterion #3 with the following: > 1 lead anywhere with > 1 mm of ST segment elevation and proportionally excessive discordant ST elevation, as defined by > 25% of the depth of the preceding S-wave.
Take home points: The Sgarbossa criteria are a useful tool when faced with a patient with chest pain who also has a LBBB. Use these rules to help you decide whether the patient’s ECG is consistent with AMI.